Characterization of Uranium Redox State in Organic-Rich Eocene Sediments

1 Characterization of uranium redox state in organic-rich Eocene sediments

3 Susan A Cumberland1,2,3, Barbara Etschmann2, Joël Brugger2, Grant Douglas4, Katy Evans5,

4 Louise Fisher6, Peter Kappen3, John W. Moreau1

5 1 School of Earth Sciences, University of Melbourne, Parkville, Victoria 3100, Australia

6 2 School of Earth, Atmosphere and Environment, Monash University, Clayton 3800, Victoria, 7 Australia

8 3 ANSTO Australian Synchrotron, 800 Blackburn Road, Clayton 3168, Victoria, Australia

9 4 CSIRO Land and Water, Floreat, Western Australia, Australia

10 5 Western Australian School of Mines, Curtin University, Bentley, Western Australia, 11 Australia

12 6 CSIRO Mineral Resources, Bentley, Western Australia, Australia

14 Abstract

15 The presence of organic matter (OM) has a profound impact on uranium (U) redox cycling,

16 either limiting or promoting the mobility of U via binding, reduction, or complexation. To

17 understand the interactions between OM and U, we characterised U oxidation state and

18 speciation in nine OM-rich sediment cores (18 samples), plus a lignite sample from the

19 Mulga Rock polymetallic deposit in Western Australia. Uranium was unevenly dispersed

20 within the analysed samples with 84% of the total U occurring in samples containing

21 >21 wt. % OM. Analyses of U speciation, including x-ray absorption spectroscopy and

22 bicarbonate extractions, revealed that U existed predominately (~71%) as U(VI), despite the

23 low pH (4.5) and nominally reducing conditions within the sediments. Furthermore, low

24 extractability by water, but high extractability by a bi-carbonate solution, indicated a strong

25 association of U with particulate OM. The unexpectedly high proportion of U(VI) relative to

26 U(IV) within the OM-rich sediments implies that OM itself does not readily reduce U, and

27 the reduction of U is not a requirement for immobilising uranium in OM-rich deposits. The

28 fact that OM can play a significant role in limiting the mobility and reduction of U(VI) in

29 sediments is important for both U-mining and remediation.

30 Keywords: Mulga Rock, uranium, mobility, organic matter, oxidation state, x-ray

31 absorption spectroscopy.

32 Highlights

33  84% of U in Mulga Rock OM-bearing sediments occurred within the OM-richest 34 samples (>21% total organic carbon; 9 of 18 samples). 35  71% of U was present in the U(VI) oxidation state. 36  Higher proportions of U occur as U(VI) in mature OM-rich sediments than previously 37 thought. 38  OM strongly complexes U(VI), limiting its mobility and reactivity. 39  U reduction processes are not ubiquitous within OM-rich sediments.

40 1 Introduction

41 The mobility of uranium (U) in natural environments is controlled by its oxidation state and

VI 2+ 42 solubility, with the uranyl ion (U O2 ) generally more soluble than U(IV). Ligands such as

2- - 3- 2- 43 CO3 , OH , PO4 and SO4 can either increase (via formation of stable aqueous complexes)

44 or decrease (via formation of insoluble minerals) U mobility (Cumberland et al., 2016).

45 Uranium mobility is also influenced by organic matter (OM) in both dissolved and

46 colloidal/particulate forms (Wood, 1996; Cumberland et al., 2016). Depending on pH,

47 dissolved OM such as humic and fulvic acids can bind U, facilitate its transportation, and

48 prevent its sorption to mineral surfaces (Luo and Gu, 2008; Zhao et al., 2012; Tinnacher et

49 al., 2013). In contrast, particulate or solid phase OM can adsorb and accumulate U, resulting

50 in the formation of OM-rich sedimentary U deposits (Greenwood et al., 2013; Cumberland

51 et al., 2016). While considerable information exists on U mobility in inorganic systems

52 (Grenthe et al., 2004), datasets for organic-rich systems are comparatively deficient (Bargar

53 et al., 2013). The complex molecular-scale associations of U and OM, and the impact of OM

54 on U mobility/immobilisation, are still being unravelled (Bargar et al., 2013). These

55 knowledge gaps limit development of in situ mining technologies (Zammit et al., 2014), as

56 well as environmental management and remediation strategies for U mining sites.

57 Recent studies have been directed towards understanding U mobility in OM-rich modern

58 sediments, including the pathways through which U potentially accumulates (Tokunaga et

59 al., 2005; Law et al., 2011). Peat, for example, can concentrate U by up to factor of 10,000

60 from groundwater containing ppb level U concentrations (e.g. (Idiz et al., 1986; Read et al.,

61 1993; Owen and Otton, 1995; Lidman et al., 2012)). Bryan et al. (2012) and Warwick et al.

62 (2005) identified U sorption to carboxylic and phenolic functional groups on insoluble

63 humics and fulvics as the likely scavenging mechanism. Furthermore, U accumulation in

64 peat and other OM-rich environments is commonly conceptualised as involving initial U(VI)

65 adsorption followed by reduction to U(IV), with the latter process resulting in long-term U

66 immobilisation (Spirakis, 1996). A suite of factors may facilitate U reduction in sediments,

2+ 67 including increased temperature (Nakashima 1992), electron donors such as H2S or Fe , and

68 direct enzymatic reduction (Newsome et al., 2014; Campbell et al., 2015). Mineral surfaces

69 (e.g., pyrite; Fe-Ti-oxides) and insoluble OM may also play roles in reduction of sorbed U(VI),

70 catalysing rate-limiting reaction steps by facilitating electron transfer from underlying

71 material or co-adsorbates (Renock et al., 2013; Latta et al., 2014). Most U in fossil economic

72 U deposits, including OM-rich types (Min et al., 2000; Deditius et al., 2008), has been

73 assumed to be associated with U(IV)-minerals, such as coffinite, or other non-crystalline

74 forms (Bhattacharyya et al., 2017). Interestingly, synchrotron X-ray absorption spectroscopy

75 (XAS) data have revealed the presence of both U(IV) and U(VI) in some OM-rich U deposits,

76 with U(IV) constituting of 35-68% of total U (Mikutta et al., 2016). Regenspurg et al. (2010)

77 found a mixture of U(VI) and U(IV) in alpine soils from the Dischma Valley (Switzerland),

78 based on combined carbonate extractions and data from X-ray Absorption Near Edge

79 Structure (XANES) analyses. These results point to U(VI) being an important component of

80 U bound to OM surfaces. From these recent findings, a picture is emerging for U-OM

81 association in modern sediments that emphasizes the complexity and variability of U redox

82 state and chemical speciation (Mikutta et al., 2016; Bone et al., 2017).

83 Synchrotron x-ray absorption spectroscopy (XAS) provides a suitable analytical technique for

84 obtaining U-oxidation state and speciation in OM-rich sediments, since it is sensitive to both

85 ‘invisible’ (e.g., colloidal; adsorbed; associated with OM, typically sub > 10 nm) and mineral-

86 bound U (Denecke et al., 1998a; Denecke et al., 1998b; Mikutta et al., 2016; Bhattacharyya

87 et al., 2017). XAS can be applied to samples with a complex composition and containing a

88 wide range of U concentrations (from a few ppm to wt.%). Here our objective was to

89 characterise the physical and chemical nature of U within an OM-rich Eocene sedimentary U

90 deposit, Mulga Rock (Western Australia), in order to determine the relative abundance and

91 origins of observed U(IV) or U(VI). Our results inform strategies for sustainable mining

92 (subsurface extraction) or environmental remediation of U in OM-rich sedimentary or

93 aqueous environments.

94 2 Geological setting and core sampling

95 The Mulga Rock (MR) U deposit is located near Kalgoorlie (Western Australia; Figure 1),

96 hosted within a series of paleochannels and lacustrine beds formed during the Eocene (56-

97 34 Ma), now buried to shallow depths (30 to 50 m). According to the petrographic analyses

98 of Jaraula et al. (2015), MR sediments contain a mixture of particulate and non-particulate

99 OM consisting of woody (lignite) material combined with aquatic algal and bacterial

100 biomass, consistent with accumulation from a combination of forested, wetland and

101 lacustrine environments. Over time, the OM layers and deeper sediments became anoxic,

102 with considerable evidence for microbial production of authigenic minerals, most

103 prominently Fe-sulphides. Uranium accumulation likely resulted from groundwater flow

104 through permeable sandstones within palaeochannels, and scavenging of the carried U

105 (present day average 8 µg L-1 U in local groundwater) by OM-rich layers (Douglas et al.,

106 2011; Jaraula et al., 2015). U-Pb isotope systematics and U-Th disequilibria studies suggest

107 that U and the associated metals originated from local lamproitic or carbonatitic sources

108 (Douglas et al., 2011; Jaraula et al., 2015).

109

110 111 Figure 1. Map of Mulga Rock, deposit sites, core locations and detail from core 5613 112 showing lithology, oxidation state and radioactivity (obtained from down-hole gamma log 113 counts (Vimy, 2014)). See also SI Table 1 for locations.

114 To date, four orebodies with potential U resources have been delineated at MR:

115 Ambassador, Emperor, Princess and Shogun (Figure 1). Ambassador, discovered in 1979, has

116 been the main subject of industry and academic investigations, with resources estimated at

117 13,000 tonnes U (Douglas et al., 1993; Douglas et al., 2011). The most recent orebody,

118 Princess, was discovered in March 2012. The wider MR deposit is estimated to contain

-1 119 66.5 MT of ore at 520 mg kg U3O8 (Vimy, 2016c). The geological setting and close

120 association of U with OM-rich sediments at MR make it an excellent site for studying the

121 nature of U in ore-grade OM-rich environments.

122 We analysed 18 samples from nine air- and diamond-drilled cores obtained from the March

123 2012 drilling project: Ambassador (4), Emperor (1), Princess (3) and Shogun (1), together

124 with a section of lignite (CD 1577) extracted during a previous exploration campaign from

125 the Ambassador deposit (Vimy, 2014, 2015, 2016a, b). Lignite sample CD 1577 was

126 immediately preserved in epoxy resin to prevent oxidation. The March 2012 samples were

127 selected on-site as they were extracted (sample locations given in Error! Reference source

128 not found.) and immediately stored in airtight containers, taking care to minimize exposure

129 to air and to leave no air gap. The samples were selected to represent a range of redox

130 states, mineralisation types and grades, as estimated on-site by a handheld Niton XL3t X-ray

131 Fluorescence (XRF) spectrometer, and on the radiometrics determined using a down-hole

132 gamma probe (AusLog, #T125, 33mm diameter) calibrated with a Cs137 source (Vimy, 2014,

133 2016b). Laboratory based geochemical analysis was completed 6-12 months after sampling

134 and preservation; synchrotron X-ray analysis was performed within two years of sample

135 preservation.

136 3 Geochemical Analysis

137 Six analytical methods were used: bulk core sample analysis (ICP-MS); chemical extractions

138 (ICP-MS); electron microscopy; and synchrotron-based XAS and X-ray fluorescence

139 microscopy (XFM).

140 3.1 Sample preparation and measurements of loss on ignition (LOI) and pH

141 Samples were characterised using fresh, oven-dried, ashed or dried-homogenised material.

142 To obtain dry masses, sub-samples and replicates were oven-dried at least overnight at

143 104 oC and weighed until constant mass. The resulting oven-dried samples either underwent

144 LOI analysis, or were homogenised to a very fine powder in an agate-lined ball mill and dry-

145 stored in glass vials at ambient temperatures until further analysis. The LOI analysis was

146 undertaken on a subset of samples (replicates = 5, n = 12) at 550 oC for 4 hr, then 900 oC for

147 2 hr, using a Thermoline muffle furnace (Dean, 1974; Heiri et al., 2001). A reduced number

148 of samples were analysed for LOI (12 of 18) to conserve material. Powdered samples were

149 analysed for total nitrogen (TN), organic carbon (TOC) and sulphur (TS) by an Elementar III

150 Elemental Analyser (University of Kiel) without further pre-treatment. Sediment pH was

151 analysed within 3 months of sample recovery using the method of (Rowell, 1994).

152 Measurements were conducted on duplicate samples using a Thermo portable meter with

153 standard calomel electrode. Six grams of undried sample was mixed with 15 ml of water, the

154 slurry was shaken by hand for 15 minutes, and the pH was then measured immediately and

155 again after 10 minutes.

156 3.2 Total and extractable metals including U

157 Total metals, U and rare earth elements (REE) concentrations were obtained from acid-

158 digested samples using a method adopted from Kamber et al. (2003) and Eggins et al.

159 (1997). Both dried and ashed samples were analysed for U (SI Table 5). 2 mL of triple-

160 distilled HF and 1 mL triple-distilled HNO3 were added to 100 mg of sample in capped Teflon

161 vessels and digested at 135 ˚C overnight on a hotplate. The dehydrated material was

162 refluxed twice with concentrated HNO3, then dissolved overnight in 3N HNO3. The solutions

163 were transferred to transparent polycarbonate tubes, and diluted with water. Solutions

164 were diluted further to a factor of 5000 with a solution of 1.6 % HNO3, then analysed using

165 inductively coupled plasma-mass spectrometry (quadrupole ICP-MS, Agilent 7700x). Results

166 of digested samples are reported as mg kg-1 of dry mass. There was a significant correlation

167 (R2 = 0.999, p < 0.005) between the ashed (n=12) and unashed samples, indicating

168 confidence in that the organic matrix did not interfere with the ICP-MS analysis. 8

- 169 Bi-carbonate (HCO3 ) (Zhou and Gu, 2005) and pure water extractions (ultra pure, 18.2 MΩ)

170 were undertaken to further constrain U speciation and oxidation state on 18 sub-samples.

171 Typically, bi-carbonate extraction assumes that in the absence of oxygen, U(VI) forms strong

172 complexes with carbonate ions and hence becomes highly soluble, whereas U(IV), including

173 unoxidised biogenic U(IV), remains uncomplexed and insoluble (Guillaumont et al., 2003;

174 Zhou and Gu, 2005). In some circumstances, monomeric (i.e., sorbed) biogenic U(IV) can be

175 released in concentrated bicarbonate solutions, without affecting uraninite/coffinite

176 stability (Alessi et al., 2012). Thus in our work the extracted U(VI) may also contain a small

177 fraction of mobile, highly reactive biogenic U(IV). The dissolved U(VI) can be separated by

178 filtration (i.e. 0.2 μm) of the aqueous phase, and the U(IV) concentration can be estimated

179 by subtraction of the amount of soluble U analysed by ICP-MS from the total U value. This

180 provides a minimum U(VI) concentration, since some U(VI) minerals, in particular silicates,

181 phosphates and vanadates are sparingly soluble (Tokunaga et al., 2012; Kanematsu et al.,

182 2014; Mehta et al., 2014). Bi-carbonate extractions were performed in duplicate under

183 anaerobic (N2) conditions as follows: 20 ml of 1 M NaHCO3 (de-gassed) was added to ~1.7 g

184 fresh hand-ground sediment; pH was recorded; and the sample was placed in 100 mL

185 crimped and sealed serum bottles (acid washed, 10% HNO3) with the headspace replaced

186 with N2 gas to avoid oxidation. Blanks were prepared as above. The suspensions were

187 rotated on an orbital shaker (150 rpm, 25 oC) continuously for one week, and then settled

188 for 10 minutes, after which they were vacuum-filtered through 0.2 μm cellulose nitrate

189 membranes (Whatman). Filtration was generally fast, with additional filters used where

190 necessary. Filtrate solutions were acidified (1% HNO3) and then analysed by ICP-MS. The pH

191 of the bi-carbonate/sediment suspension was 8.5 ± 0.1. Water extractions were performed

192 under the same conditions but filtered using a standard 1.2 µm GF/C Whatman glass filter

193 paper to include a potentially mobile U colloidal fraction.

194 3.3 Imaging U distribution and elemental association using XFM and SEM

195 To determine the microscale spatial distribution and elemental/mineralogical association of

196 U, subsamples of unground grains from core NNA 5613 (Ambassador ore body) were

197 selected based on total activity (measured by Geiger counter) and mounted onto 0.5 cm2

198 SiN windows using a 5% glucose solution. Due to the finely dispersed nature of the U in the

199 MR samples, TEM techniques proved unsuccessful in detecting U in the samples. One

200 sample, MR5613a, yielded better SEM results and gave higher radioactive counts per

201 second, likely due to higher U (5000 mg kg-1) concentrations. This sample was therefore

202 favoured for subsequent micro-analytical characterisation.

203 Subsample grains from MR5613a were analysed at the XFM Beamline (Australian

204 Synchrotron) at 18.5 keV incident photon energy and 2 m beam size (Paterson et al.,

205 2011). Fluorescence signals were collected using a 384-pixel Maia detector (Ryan et al.,

206 2014). The fluorescence spectra collected were processed using the GeoPIXE software to

207 produce semi-quantitative elemental maps (Ryan and Jamieson, 1993; Ryan, 2000;

208 Etschmann et al., 2010; Li et al., 2016).

209 Backscatter electron images (sample MR5613a from core NNA 5613) were obtained on thin

210 sections using a Philips FEI XL30 environmental scanning electron microscope (ESEM)

211 equipped with an OXFORD INCA energy-dispersive X-ray spectrometer (EDS) (Earth Sciences,

212 University of Melbourne), and EDS maps were collected with a Field Emission Philips XL30

213 ESEM (BIO21, University of Melbourne).

214 The mineralogy of lignite sample CD 1577 was characterised using μ-XRD. Spectra were

215 collected using a Bruker general area detector diffraction system (GADDS) with Cu Kα

216 radiation and 200 or 300 µm collimators (CSIRO, Clayton, Melbourne; SI Figure 1).

217 3.4 Uranium speciation via XANES and EXAFS spectroscopy.

218 X-ray absorption near edge structure (XANES) and extended x-ray absorption fine structure

219 (EXAFS) spectra were collected at the U-L3 absorption edge at the XAS Beamline (Australian

220 Synchrotron). Spectra were collected for eight powdered core samples (MR5076a,

221 MR5076b, MR5077, MR5613a, MR5613b, MR5766b, MR5766c, MR5766d,) selected because

222 they had sufficient U concentrations (from ICP-MS results) to yield good quality spectra, and

223 nine U-bearing minerals provided by the Melbourne Museum serving as standards. Data on

224 these standards are presented in the supplementary information (Error! Reference source

225 not found., SI Figure 2 and SI Figure 3). Homogenised powdered samples or mineral

226 standards diluted with boron nitride were packed into Al holders, and sealed with Kapton

227 tape. For analysis, samples were loaded into a He cryostat (pulse-tube Optistat, Oxford

228 Instruments; T~10K). In addition, spectra were taken from a fresh section of the core

229 subsample CD 1577 at six random points. The incident energy was controlled using a Si(111)

230 double crystal monochronomator and calibrated at the Zr-K absorption edge using a Zr

231 metal foil (first maximum of the first derivative at 17998 eV). Repeat measurements of this

232 Zr foil established that there was no measurable energy shift during the experiments.

233 Uranium was detected concurrently in fluorescence (100 element Canberra HP-Ge

234 fluorescence detector) and transmission modes using ion chambers (Oken, N2 flow at ~0.3

235 L/min, U = 2.1 kV). Aluminium foil was used to reduce parasitic fluorescence count rate from

236 other fluorescence. Radiation induced changes to U valence or chemistry, assessed over

237 repeat XANES runs, were negligible. The XANES scans were acquired as follows: baseline, 11

238 17010 to 17141 eV in 3 eV steps; over the edge-step, 17144 to 17204 eV in 0.3 eV steps; and

239 with constant steps in k-space (k = 0.035 Å-1) above the edge. For EXAFS measurements the

240 energy range was extended to ~18140 eV (k~16 A-1).

241 Spectra normalisation and linear combination fitting (LCF) were performed using the

242 DEMETER software package (Ravel and Newville, 2005). Most spectra were truncated after

243 17969 eV (14 Å-1) to avoid interference from the Zr K edge. The LCF was performed using

244 two reference spectra, synthetic uraninite for U(IV) (Tsarev et al., 2016) and uranopilite for

245 U(VI) (see also SI Figure 2 ). The uranium oxidation state was determined from the LCF data.

246 The EXAFS data were processed and refined using the HORAE package (Ravel and Newville,

247 2005) with the theoretical standards calculated by FEFF9 (Rehr et al., 2010). The kn-

2 248 weighted data (n = 1, 2, 3) were refined in R space. The amplitude-reduction factor (S0 ) was

249 estimated to be 0.85 by fitting UO2(NO3)2•6H2O (structure confirmed by powder XRD; SI

250 Table 3 and Error! Reference source not found.); this value was used in the refinement of all

251 the sample data.

252 4 Results

253 4.1 Core geochemistry

254 Core sample appearance varied from sandy and dry, to dark, peaty and moist, with a mean

255 moisture content of ~12 wt. % (range 0.3 to 38 wt. %). Mean sample pH was 4.5 (range 3.5

256 to 7). The physicochemical data are presented in SI Table 4. The TOC concentrations of the

257 samples averaged 22.5 wt. % (range 1 to 59 wt. %), reflecting substantial vertical in situ

258 variation in the MR deposit. Samples LOI550 ranged from 1 to 57 wt. %, and correlated with

2 o 259 TOC (R = 0.922, slope = 1.04, p < 0.005; n = 10). Further LOI performed at 900 C (LOI900)

260 ranged from 1 to 78.3 wt. % (average = 31 wt. %). This mass loss between 550 and 900 oC

261 can be attributed to decomposition of inorganic carbonates (Dean, 1974; Heiri et al., 2001).

262

263

264

265 Figure 2. Uranium distribution and correlations with organic matter and trace elements. 266 (A) Bar chart showing uranium distribution in samples according to carbon content. The 267 data (n = 18) are split on the basis of their carbon contents (TOC) into two equal groups 268 (n = 9 for each group); one group has < 21 wt. % TOC, the other 22-60 wt. % TOC. Error bars 269 are standard deviations of uranium concentration: samples containing < 21% TOC contain 270 1600 mg U kg-1 and those with 22-60 wt. % TOC 4500 mg U kg-1. (B) Pearson correlation 271 coefficients for all core samples (n = 18); (B-1) between TOC and metals; and (B-2) U and 272 metals, where height of bar corresponds to R2 values. Stars indicate statistical significance: 273 *** p < 0.005; ** p < 0.05; * p < 0.5, ns (no star) p > 0.51. Note that Fe was not analysed in 274 these samples. (C and D) Detail of sediment grains from the MR core NNA 5613: (C- 275 1) Scanning Electron Microscope (SEM) image; (C-2) is enlargement of (C-1), showing U-rich 276 and Zn-rich particles of similar sizes. (D) Synchrotron x-ray fluorescence (XFM) image 277 showing U, Cu and Sr distribution in representative grain from the same sample as (C); U = 278 red, Cu = green and Sr = blue.

279 [full colour figure in print ] 280

281 Uranium in the core samples ranged from 2 to 15,000 mg kg-1 U (mean = 2800 mg kg-1, n =

282 18), and correlated with chemical digests performed on ashed samples (900 oC, n = 12)

283 when normalised to mg kg-1 dry mass (R2 = 0.998, slope = 1.00, p < 0.0005; Pearson) (see SI

284 Table 5). Core samples contained average element concentrations in the order

285 Ti>U>Zn>Ca>total REE>Cu>Ni>Pb>Cr>Co>Sc>V>As>Th (SI Table 6). Correlations between U

286 and TOC, and U and other metals were weak (p < 0.5), with the strongest relationships

287 occurring between U and Pb, U and total REEs, Cd and As (p < 0.005, R2 values shown in

288 Figure 2B and SI Table 7). Slight correlations were observed between TOC and Ca, Sr or V (p

289 < 0.05), but most elements show no significant correlation with TOC. Detailed analyses on

290 the organic chemistry of samples from the Mulga Rock deposit are presented elsewhere

291 (Jaraula et al 2015).

292 The spatial distribution of U in the core samples was highly variable, with U present either in

293 diffuse form (no identifiable minerals > 10 nm) or as U-rich micro- to nano- particles (see

294 XFM and SEM images in Figure 2 C and D). Sample MR5613a (from core NNA 5613, Figure 2

295 D), with a bulk U concentration of 5,000 mg kg-1, showed substantial variation in the form

296 and distribution of U between the four subsamples scanned. Two subsamples contained

297 disseminated U, while U was present as discrete µm-sized particles in the other two. The

298 lignite sample CD 1577 was structurally more coherent than the March 2012 cores. The

299 mineral matrix consisted mainly of quartz (see Error! Reference source not found. for µXRD

300 analysis). SEM elemental mapping and BSE imaging (Figures 3, 4 and SI Figure 4 for EDS

301 spectrum) show that U exists in particulate form (µm-sized), but also forms large (up to

302 >1 mm) aggregates around Fe-sulfide minerals. The U-pyrite association is common at MR

303 (Douglas et al., (2011) as well as in many sandstone-hosted U deposits (e.g., Bonnetti et al.

304 (2015; 2017); Wülser et al., (2011)), and probably reflects an association between bacterial

305 sulphate reduction and U reduction (Bonnetti et al., 2017). Micro-XRD data confirm the

306 predominance of pyrite, and show that the main U-phase surrounding the pyrite

307 grains/aggregates is coffinite [USiO4] (Error! Reference source not found.). The MR coffinite

308 has unit cell parameters (a = b = 6.943(2) Å; c = 6.262(2) Å; V = 301.9(3) Å3) that are close to

309 those of the Tertiary sandstone hosted Beverly deposit in South Australia (a = b = 6.971(2) Å;

310 c = 6.255(2) Å; V = 304.0(3) Å3; Wülser et al. (2011)).

311 312 Figure 3. Chemical maps (U, Al, C, O, Na, S and Si) based on SEM-Energy Dispersive 313 Spectrometry (EDS) of lignite sample CD 1577. Elemental concentrations are represented by 314 intensity, where bright colours show high concentrations and dark or black areas represent 315 areas where elements are absent or at low concentrations. Sulfur corresponds to pyrite

316 distribution. Si is associated with U (coffinite according to XRD results), bright spots 317 correspond to quartz. Scale bar is 100 µm. 318 [full colour figure in print] 319

320 321 Figure 4. Composite ESEM backscatter image, showing distribution of uranium in the lignite 322 section from core CD 1577. Uranium is present throughout the whole sample with the 323 brightest phase illustrating the highest U concentration. The EXAFS spectra were taken at 324 random points which showed that U was present in both U(IV) and U(VI) oxidation states 325 across the sample. Inserts (B) and (C) are enlargements of indicated areas in (A). 326

327 4.2 Leach tests and XANES linear combination fits

328 In all samples analysed, U was present in both U(IV) and U(VI) oxidation states, with usually

329 more U(VI) than U(IV). Bicarbonate extraction of U (UCO3), a process conducted under

330 anaerobic (N2) conditions which targets leachable U(VI) ions, indicated that an average of

331 69% (SD = ±18, n = 18) of total U is in the U(VI) form. The oxidation state of U (U(IV) or U(VI))

332 and total concentrations of U are highly correlated (R2 = 0.996, p < 0.005) with the

333 percentage of U(VI) increasing with total U concentration, giving a slope of 0.79 for U(VI)

334 versus total U (Figure 5A). Figure 5B shows that samples with low TOC (<8%) have a highly

335 variable oxidation state, whereas samples with high TOC (i.e. > 20%) tend to have a

336 relatively constant oxidation state, dominated by U(VI) (75±7% U(VI)). Interestingly,

337 extractions using only deionised (ultra-pure, 18.2 MΩ) water under the same anaerobic

338 conditions as bi-carbonate method, only removed 2.5 % of the total U, showing that the

339 majority of U was not readily water-soluble.

340 The U oxidation state (UXAS) was independently derived from the LCF results as described

341 above (see also Error! Reference source not found. and Figure 5 C and D for the normalized

342 XANES spectra). The LCF data (Error! Reference source not found.) show that the uranium is

343 predominately present (i.e. 74 %) in the higher oxidation state, +6. These results are

344 consistent with the concentration values from CO3-extraction. Comparison of XAS (74%) and

345 bi-carbonate extraction datasets (69%) showed good agreement, giving an overall mean

346 across the two methods of 71% U(VI).

347

348

349 Figure 5. Uranium speciation as derived from bicarbonate extraction (A and B); and 350 synchrotron-XANES analyses (C and D). (A) Bi-carbonate extracted U(VI) and U(IV) data in 351 mg kg-1 (y axis) plotted as a function of total U (mg kg-1). Black squares are U(VI) (R² = 0.996. 352 y = 0.7896x – 49.5 for U(VI)) and red dots are U(IV) (R2 =0.21 x 49.5). (B) U(IV) and U(VI) 353 oxidation states as percent of total U (U(VI)/(U(VI)+U(IV)) x 100 plotted as a function of TOC 354 wt. %. Black squares =U(VI) ,and red dots =U(IV). (C) XAS-XANES spectra of MR bulk core 355 samples and the lignite sample CD1577. The overlain dotted line is the fitted spectra from 356 LCF The arrow marks the position of the O=U=O U(VI) shoulder. (D) are the derivatives of 357 (C).

358

359 4.3 EXAFS fits

360 Two types of XANES spectra were obtained from six measurements performed on the lignite

361 sample CD 1577 (Figure 4). These spectra were grouped, merged and analysed separately,

362 with one type showing a uranyl-like (‘oxidised-CD1577’) spectrum and the other a reduced

363 (‘reduced-CD1577) U spectrum (Figure 6). The EXAFS data also differ greatly among the two

364 types (Figure 6), and different fitting strategies were used for each type. EXAFS data from

365 the samples MR5766c and MR5613a are similar to the ‘oxidised-CD1577’ lignite spectra, and

366 were refined simultaneously.

367 4.3.1 Fitting of the EXAFS spectra with uranyl-like XANES

368 EXAFS spectra from MR5766c, MR5613a and ‘oxidised-CD1577’ were analysed with a

369 common E0, using selected paths from the uranopilite structure (amcsd 0005730; Burns

370 (2001)), and one uranium-carbon single-scattering path from the structure of C15H10O8U

371 from de Lill and Chan (2013). The advantage of this approach is that the ‘oxidised-CD1577’

372 spectra could be fitted by constraining the distances and Debye-Waller factors to be the

373 same as those for MR5766c. To obtain the best fit for the ‘oxidised-CD1577’ spectra, the

374 number of oxygen atoms (O15 only) was refined, allowing additional oxygen to be fitted

375 that was not present in the MR5766c and MR5613a spectra. Note that the choice of C rather

376 than another light element (e.g. O, N, P) cannot be justified by EXAFS alone. Carbon was

377 chosen because the refined bond-distance and coordination are consistent with previous

378 studies (see below); SEM did not reveal a correlation between U and P, or any of the poorly

379 soluble uranyl phosphates minerals; and the samples are rich in C but poor in N (C/N≥100;

380 Jaraula et al. 2015). The final results are shown in Table 1.

381 Two models were used for comparison: a simple model where the two equally short U-O

382 distances (uranyl ion) were constrained to be the same (Table 1); and a second model where

383 the two short U-O distances were refined independently (SI Table 9). While more complex

384 multiple scattering (MS) paths (U-Oax2-U-Oax2, U-Oax2-Oax1 and U-Oax2-U-Oax1) could be

385 employed to fit the data out to a distance of 3.58 Å in the second model, the two short U-O

386 bonds were the same within error (1.75(3) and 1.77(3) Å). Furthermore, the goodness-of-fit

387 parameters for both models were effectively identical, indicating that there was no

388 significant advantage to the second model, so that the simpler model (Table 1) is preferred.

389 In all cases, the fit resulted in two carbon atoms at distances of 2.90(7) to 2.93(3) Å. This

390 configuration suggests a bidentate coordination to functional groups of the solid OM. All

391 U(VI)-dominant spectra lacked a peak at distance of 3.8 Å, which would indicate the

392 presence of U-U bonds. This shows that most U is not present in the form of uranyl

393 minerals. Consequently, the most U(VI) in the studied samples is present predominantly as a

394 monomeric U(VI) complex bound to OM. This is further supported by the flattening of the

395 oscillation between 6.8 and 7.9 Å-1 in the k2-weighted EXAFS data (arrow, Figure 6A), a

396 feature which appears in uranyl-humic compounds (Denecke et al., 1998a; Denecke et al.,

397 1998b; Mikutta et al., 2016).

398

399 Table 1 Fitting parameters obtained from U L3-Edge of EXAFS Spectra. Ntot = k-range 2 2 -1 2 sample ligand N R (A) ss (A ) fraction N*fraction ΔE0 red X r-factor (Å ) R-range (A) k-weighting S0 Oax2 2 1.773(5) 0.003 (fix) 9.2(5) 113 0.023 2 - 12 1.3 - 4 1,2,3 0.85 MR5766c O15 2 2.22(1) 0.003 (fix) O14 2 2.36(1) 0.003 (fix) OH20 2 2.42(2) 0.003 (fix) C 2 2.93(3) 0.003 (fix) U-Oax2-U-Oax2 (MS) 3.545(7) 0.006 (fix)

Oax2 2 1.77(2) 0.003 (fix) 11(3) 113 3 - 11.5 1.3 - 4 1,2,3 0.85 MR5613 a O15 2 2.24(4) 0.003 (fix) O14 2 2.39(4) 0.003 (fix) OH20 2 2.45(6) 0.003 (fix) C 2 2.90(7) 0.003 (fix) U-Oax2-U-Oax2 (MS) 3.55(3) 0.006 (fix)

oxidised- CD1577* Oax2 2.3(2) 1.773(5) 0.003 (fix) 9(1) 113 2.5 - 12 1 - 4 1,2,3 0.85 O15 2.8(4) 2.22(1) 0.003 (fix) O14 2 2.36(1) 0.003 (fix) OH20 2 2.41(2) 0.003 (fix) C 2 2.91(4) 0.003 (fix) U-Oax2-U-Oax2 (MS) 3.545(7) 0.006 (fix)

Reduced-CD1577 Oax2 2** 1.74(2) 0.003 (fix) 0.47(10) 0.94 8(3) 199 3 – 10.5 1 - 4 1,2,3 0.85 O15 2** 2.35(7) 0.003 (fix) 0.47(10) 0.94 O14 2** 2.33 (fix) 0.003 (fix) 0.47(10) 0.94 OH20 2** 2.37 (fix) 0.94

O1 (coffinite) 2** 2.25(3) 0.003 (fix) 0.53(10) 1.06 O2 (coffinite) 4** 2.83(3) 0.003 (fix) 0.53(10) 2.12 U1 (coffinite) 4** 3.79(4) 0.005(4) 0.53(10) 2.12 400 * These distances were constrained to be the same as those of MR5766c and were fitted together. 401 ** The number of ligands was fixed, the fraction of each mineral component was refined.

402 403

404 Figure 6 EXAFS data and fits, shown in (A) k-space; (B) R–space; (C) R- Imaginary. The uranyl- 405 like spectra are shown at the top of each graph for MR5766c, MR5613a and CD1577 406 oxidised spectra (ox-MR) with relevant paths shown below. The reduced spectra (red-MR) 407 from CD1577 is shown below in Red and relevant paths underneath.

408

409 4.3.2 Fitting of the reduced-CD1577 EXAFS spectra

410 The spectrum of the reduced component from the lignite (reduced-MR, CD 1577) was fitted

411 using a wide variety of models. We report here a simple model (SI Table 3) based on a

412 combination of paths corresponding to uranopilite [(UO2)6(SO4)O2(OH)6(H2O)6·8H2O], as a

413 proxy for the uranyl component (amcsd 0005730; (Burns, 2001); and coffinite [USiO4] (icsd

414 15484; (Fuchs and Gebert, 1958), which is the main U mineral according to micro-XRD (SI

415 Figure 1). The best fit was obtained using 47(10) % uranopilite and 53(9.8) % coffinite. The

416 presence of atoms at ~3.8 Å distance indicates a second shell of U consistent with the crystal

417 structure of coffinite (EXAFS fit U-U distance of 3.79(4); crystallographic distance of

418 3.83(2) Å). Furthermore, Bhattacharyya et al. (2017) report a U-C distance of 2.90 Å for their 23

419 organic U(IV) structure, but this contribution is absent from the reduced-MR sample. Thus,

420 in contrast to U(VI), which is present in mostly monomeric form within the organic matrix,

421 U(IV) is present in mineral form (predominantly coffinite in CD 1577).

422 5 Discussion

423 5.1 U-OM association at MR

424 Accumulation of U by OM by 10’s of 1000’s of times relative to U-concentrations in

425 groundwater is a common and long established feature. Analyses of groundwater by

426 Douglas et al. (2011) show that modern waters at MR contain ppb concentrations of U

427 (8 µg L-1), hence implying concentrations of up to > 6 orders of magnitude in the OM layers

428 (mean = 2500 mg kg-1 U, max = 15,000 mg kg-1 U). The highest concentrations of U at MR

429 were associated with high concentrations of organic carbon, with 84% of the total U in half

430 the sample set (n=9) containing >22-60 wt. % TOC. According to Douglas et al. (2011),

431 uranium and the associated metals precipitated syngenetically with OM as it was deposited

432 during a humid phase in the Late Eocene. Subsequent small-scale mobilisation during

433 diagenesis and/or weathering concentrated the metals in the upper 2 m of the lignite, with

434 the latest episode in the last 300,000 years. Continuous upgrading via interaction with

435 groundwater containing low concentrations of dissolved U could also have contributed to

436 the high grades observed today.

437 A characteristic of the OM-rich layers is that the proportion of U(VI) is relatively consistent

438 (e.g. 77.2% ± 18 from extractions), despite the fact that absolute U concentrations vary

439 greatly. Correlation coefficients between U and OM in soils and sediments in the literature

440 are highly variable (Meunier et al., 1989; Landais, 1996; Min et al., 2000; Regenspurg et al.,

441 2010; Cumberland et al., 2016; Och et al., 2016) and Mulga Rock is no exception. At MR the

442 OM-rich samples contained most of the U identified across all the studied samples, yet no

443 statistically significant correlation existed between grade and OM content. We infer that U

444 introduced in oxidized form (uranyl complexes) by groundwater flowing through the OM-

445 rich sediments sorbed strongly and immediately to the available sites on the organic carbon.

446 We examined MR samples for U oxidation state, as this parameter controls the potential for

447 U mobility and biotoxicity, as well as providing necessary data to inform design of the ore-

- 448 forming process. The results of both XAS and HCO3 -chemical extractions show that on

449 average 71% of the total U (mean of 69% HCO3 extraction and 74% XAS LCF) was in the U(VI)

450 oxidation state in the MR samples. That most of the U was observed as U(VI) was surprising

451 given the long-held assumption that for long-term U immobilisation and accumulation to

452 occur within OM-rich sediments, reduction would be more important than adsorption

453 (Spirakis, 1996). In the case of MR specifically, the presence of high proportions of reduced

454 U might also have been expected, given the low to neutral pH (3.5 to 7), reducing

455 conditions, and abundant coffinite/uraninite particles within the ores (Douglas et al., 2011).

456 However, while U phase stability modelling at low pH and Eh predicts that uraninite (UO2)

457 and coffinite (under quartz-supersaturation conditions; Brugger et al., (2005)) are the

458 dominant phases in the absence of OM (Langmuir, 1978), models that include OM as

459 humate suggest a U(VI)-humate phase dominating at ~pH 6 (Shanbhag and Choppin, 1981;

460 Lenhart et al., 1997; Cumberland et al., 2016).

461 Recent studies have shown that modern OM-rich soils contain a high proportion of U(VI) (32

462 -65 % U(VI) in peats, and 51 to 100 % U(VI) in organic-rich alpine soils; (Regenspurg et al.,

463 2010; Mikutta et al., 2016). Hence, the predominance of U(VI) at MR might be explained by

464 short U residence times and slow reduction kinetics. However, Jaraula et al. (2015) showed

465 that U has been present at MR long enough to cause radiolytic damage to organic

466 biomarkers. We therefore interpret that since the first sediments were deposited, U(VI)

467 reduction has not played a major role in immobilizing U, despite the fact that MR is a

468 shallow (<50 m) deposit undergoing deep weathering, and has remained in contact with

469 groundwater containing µg L-1 levels of U. Furthermore, the ores exist in an environment

470 where both inorganic and microbial redox processes are expected to be promoted – as

471 indeed demonstrated locally by the abundance of biogenic sulphide minerals.

472 5.2 Uranium - organic matter complexes

473 The poor capability of pure water to extract U compared to the bicarbonate leach (2.5%

474 versus 69% extraction) suggests poor U(VI) solubility and strong U(VI) association to

475 particulate OM (i.e. > 1.2 µm filter-size) or possibly finely dispersed silica. We therefore

476 postulate that particulate OM has helped to retard U mobility at MR, especially since only

477 very low concentrations of U persisted within neighbouring sandstone sediments that were

478 lower in organic carbon.

479 Mikutta et al. (2016) reported that uranyl present in modern peats (e.g. <13,000 years) is

480 surrounded by 0.9 to 2.0 carbons at 2.89-2.93 Å. Therefore the speciation in the MR lignite,

481 where we identified a bidentate C coordination at similar distance (2.90(7) to 2.93(3) Å)

482 appears to be surprisingly similar to that in modern peat. This U-carbon model is also

2+ 483 consistent with that of Denecke et al. (1998a; 1998b), representing poorly-ordered UO2

484 bonding to COOH- (carboxylic acid) groups present in organic macromolecules (Kaplan et al.,

485 2016). Such bonding is likely at MR, considering that Douglas et al. (2011) found that up to

486 13.1% of the carbon is present in carboxyl groups within the MR OM-rich sediments.

487 Bhattacharyya et al. (2017) showed that uranium could remain attached to these carboxyl

488 ligands even if U was reduced in situ to U(IV); however at MR, U reduction appears to be

489 mainly associated with the formation of coffinite and/or uraninite particles.

490 6 Conclusion

491 Our findings show that uranium at Mulga Rock was mostly (85%) associated with the

492 organic-rich sediments, where most of it is present in the oxidised state of U(VI) (71% of

493 total U). It is commonly assumed that the flow of groundwater largely determines the

494 distribution of the U; however, this study found that U was associated with higher organic

495 matter concentrations, and therefore the mobility of U can be inhibited by OM. Results

496 from EXAFS analyses on the oxidised CD1577 spectra are consistent with the interpretation

497 that uranium predominantly forms a bidendate U(VI) complex with two organic (likely

498 carboxyl) groups.

499 Despite the dominant presence of OM, U(VI) was found to be reduced only incompletely,

500 even under low pH conditions. Hence, MR-OM itself did not readily reduce U, and reduction

501 of U was not required for immobilisation of uranium over geological time frames as

502 applicable to Mulga Rock. Tetravalent U, which displayed consistency with a coffinite–like

503 structure, was found to be widely present at Mulga Rock; these species most likely result

504 from in situ U(VI) reduction as a consequence of heterogeneous secondary processes, such

505 as biomediated sulfate reduction and pyrite mineralisation and/or microbial enzymatic

506 activity.

507 Our results extend a growing body of evidence that U speciation in OM-rich geological

508 materials involves a complex mixture of U(IV) and U(VI), representing potentially

509 overprinting generations of U delivery by groundwater and subsequent biologically-

510 mediated U(VI) reduction. These results have direct implications for the development of U-

511 OM chemical extraction methods and post-mining management practices: In the context of

512 in-situ recovery, oxidation is not necessarily required for mobilising U, as U(VI) is present in

513 higher proportions than previously assumed. In the case of remediation, environmental

514 management strategies can shift from a focus on U reduction to using OM to capture U in its

515 uranyl form, as a potentially equally effective or preliminary approach to U remediation.

516 Future research into different forms of environmental U-OM interactions, in the presence of

517 other potentially reactive components such as mineral surfaces and microorganisms, is

518 necessary, to understand the conditions under which U can be released or immobilised in

519 OM-rich waters and sediments.

520 7 Acknowledgements

521 This work was funded through a grant to JWM from the CSIRO Organic Geochemistry of

522 Mineral Systems (OGMS) Research Cluster. Parts of this research were undertaken on the

523 XAS and XFM beamline at the Australian Synchrotron, part of ANSTO. We would like to

524 thank: Xavier Moreau (Vimy Resources) for MR samples; Dr Alan Grieg, Graham Hutchinson

525 and Dr Hong Vu for laboratory assistance; Prof. Lorenz Schwark (University of Kiel) for TOC,

526 N and S data analysis; Dr Martin De Jonge and Dr Daryl Howard for beamline assistance at

527 AS XFM, Museum Victoria for comparative minerals, and Steven Henderson for his editorial

528 contributions. SC would like to thank the Australian Synchrotron (AS) for beamtime and the

529 AS and Monash University for support in preparation of this manuscript through a joint

530 post-doctoral fellowship.

531

532 8 References

533

534 Alessi, D.S., Uster, B., Veeramani, H., Suvorova, E.I., Lezama-Pacheco, J.S., Stubbs, J.E., Bargar, J.R., 535 Bernier-Latmani, R., 2012. Quantitative Separation of Monomeric U(IV) from UO2 in Products of 536 U(VI) Reduction. Environmental Science & Technology 46, 6150-6157, doi:10.1021/es204123z. 537 Bargar, J.R., Williams, K.H., Campbell, K.M., Long, P.E., Stubbs, J.E., Suvorova, E.I., Lezama-Pacheco, 538 J.S., Alessi, D.S., Stylo, M., Webb, S.M., Davis, J.A., Giammar, D.E., Blue, L.Y., Bernier-Latmani, R., 539 2013. Uranium redox transition pathways in acetate-amended sediments. Proceedings of the 540 National Academy of Sciences 110, 4506-4511, doi:10.1073/pnas.1219198110. 541 Bhattacharyya, A., Campbell, K.M., Kelly, S.D., Roebbert, Y., Weyer, S., Bernier-Latmani, R., Borch, T., 542 2017. Biogenic non-crystalline U(IV) revealed as major component in uranium ore deposits. Nature 543 Communications 8, 15538, doi:10.1038/ncomms15538. 544 Bone, S.E., Dynes, J.J., Cliff, J., Bargar, J.R., 2017. Uranium(IV) adsorption by natural organic matter in 545 anoxic sediments. Proceedings of the National Academy of Sciences 114, 711–716, 546 doi:10.1073/pnas.1611918114. 547 Bonnetti, C., Cuney, M., Michels, R., Truche, L., Malartre, F., Liu, X., Yang, J., 2015. The Multiple Roles 548 of Sulfate-Reducing Bacteria and Fe-Ti Oxides in the Genesis of the Bayinwula Roll Front-Type 549 Uranium Deposit, Erlian Basin, NE China. Economic Geology 110, 1059-1081. 550 Bonnetti, C., Liu, X., Zhaobin, Y., Cuney, M., Michels, R., Malartre, F., Mercadier, J., Cai, J., 2017. 551 Coupled uranium mineralisation and bacterial sulphate reduction for the genesis of the Baxingtu 552 sandstone-hosted U deposit, SW Songliao Basin, NE China. Ore Geology Reviews 82, 108-129, 553 doi:10.1016/j.oregeorev.2016.11.013. 554 Brugger, J., Long, N., McPhail, D.C., Plimer, I., 2005. An active amagmatic hydrothermal system: The 555 Paralana hot springs, Northern Flinders Ranges, South Australia. Chemical Geology 222, 35-64, 556 doi:10.1016/j.chemgeo.2005.06.007. 557 Bryan, N.D., Abrahamsen, L., Evans, N., Warwick, P., Buckau, G., Weng, L., Van Riemsdijk, W.H., 558 2012. The effects of humic substances on the transport of radionuclides: Recent improvements in 559 the prediction of behaviour and the understanding of mechanisms. Applied Geochemistry 27, 378- 560 389, doi:10.1016/j.apgeochem.2011.09.008. 561 Burns, P.C., 2001. A new uranyl sulfate chain in the structure of uranopilite. The Canadian 562 Mineralogist 39, 1139-1146. 563 Campbell, K.M., Gallegos, T.J., Landa, E.R., 2015. Biogeochemical aspects of uranium mineralization, 564 mining, milling, and remediation. Applied Geochemistry 57, 206-235, 565 doi:10.1016/j.apgeochem.2014.07.022. 566 Cumberland, S.A., Douglas, G., Grice, K., Moreau, J.W., 2016. Uranium mobility in organic matter-rich 567 sediments: A review of geological and geochemical processes. Earth-Science Reviews 159, 160-185, 568 doi:10.1016/j.earscirev.2016.05.010. 569 de Lill, D.T., Chan, B.C., 2013. Structure and luminescence of a one-dimensional uranium 570 coordination polymer assembled through benzophenone-4,4′-dicarboxylate. Inorganica Chimica Acta 571 404, 215-218, doi:https://doi.org/10.1016/j.ica.2013.03.006. 572 Dean, W.E., 1974. Determination of carbonate and organic matter in calcareous sedimentry rocks by 573 loss of ignition:Comparison with other methods. Journal of Sedimentary Petrology 44, 242-248, 574 doi:doi:10.1306/74D729D2-2B21-11D7-8648000102C1865D 575 Deditius, A.P., Utsunomiya, S., Ewing, R.C., 2008. The chemical stability of coffinite, USiO4nH(2)O; 576 0

580 with organic acids using EXAFS. Journal of Alloys and Compounds 271, 123-127, doi:10.1016/s0925- 581 8388(98)00038-3. 582 Denecke, M.A., Reich, T., Pompe, S., Bubner, M., Heise, K.H., Nitsche, H., Allen, P.G., Bucher, J.J., 583 Edelstein, N.M., Shuh, D.K., Czerwinski, K.R., 1998b. EXAFS investigations of the interaction of humic 584 acids and model compounds with uranyl cations in solid complexes. Radiochimica Acta 82, 103-108. 585 Douglas, G.B., Butt, C.R.M., Gray, D.J., 2011. Geology, geochemistry and mineralogy of the lignite- 586 hosted Ambassador palaeochannel uranium and multi-element deposit, Gunbarrel Basin, Western 587 Australia. Mineralium Deposita 46, 761-787, doi:10.1007/s00126-011-0349-4. 588 Douglas, G.B., Gray, D.J., Butt, C.M., 1993. Geochemistry, minerology and hydrogeochemistry of the 589 Ambassador multi-element lignite deposit, Western Australia: With additional investigations on the 590 characterization of organic matter. CSIRO, Australia. 591 Eggins, S.M., Woodhead, J.D., Kinsley, L.P.J., Mortimer, G.E., Sylvester, P., McCulloch, M.T., Hergt, 592 J.M., Handler, M.R., 1997. A simple method for the precise determination of ≥ 40 trace elements in 593 geological samples by ICPMS using enriched isotope internal standardisation. Chemical Geology 134, 594 311-326, doi:https://doi.org/10.1016/S0009-2541(96)00100-3. 595 Etschmann, B.E., Ryan, C.G., Brugger, J., Kirkham, R., Hough, R.M., Moorhead, G., Siddons, D.P., De 596 Geronimo, G., Kuczewski, A., Dunn, P., Paterson, D., de Jonge, M.D., Howard, D.L., Davey, P., Jensen, 597 M., 2010. Reduced As components in highly oxidized environments: Evidence from full spectral 598 XANES imaging using the Maia massively parallel detector. American Mineralogist 95, 884-887, 599 doi:10.2138/am.2010.3469. 600 Fuchs, L.H., Gebert, E., 1958. X-Ray Studies of Synthetic Coffinite, Thorite and Uranothorites. 601 American Mineralogist 43, 243-248. 602 Greenwood, P., Brocks, J., Grice, K., Schwark, L., Jaraula, C., Dick, J., Evans, K., 2013. Organic 603 Geochemistry and Mineralogy: I. Characterisation of Organic Matter Associated with Metal Deposits. 604 Ore Geology Reviews, 1-27, doi:10.1016/j.oregeorev.2012.10.004. 605 Grenthe, I., Fuger, J., Konings, R.J.M., Lemire, R.J., Muller, A.B., Nguyen-Trung, C., Wanner, H., 2004. 606 Chemical Thermodynamics of Uranium. OECD- NEA, Paris, France. 607 Guillaumont, R., Fanghanel, T., Neck, V., Fuger, J., Palmer, D., Grenthe, I., Rand, M.H., 2003. Update 608 on the chemical thermodynamics of uranium, neptunium, plutonium, americium, and technetium. 609 OECD, Nuclear Energy Agency, Paris, France, p. 959. 610 Heiri, O., Lotter, A., Lemcke, G., 2001. Loss on ignition as a method for estimating organic and 611 carbonate content in sediments: reproducibility and comparability of results. Journal of 612 Paleolimnology 25, 101-110, doi:10.1023/a:1008119611481. 613 Idiz, E.F., Carlisle, D., Kaplan, I.R., 1986. Interaction between organic matter and trace metals in a 614 uranium rich bog, Kern County, California, U.S.A. Applied Geochemistry 1, 573-590, 615 doi:10.1016/0883-2927(86)90065-X. 616 Jaraula, C.M.B., Schwark, L., Moreau, X., Pickel, W., Bagas, L., Grice, K., 2015. Radiolytic alteration of 617 biopolymers in the Mulga Rock uranium deposit. Applied Geochemistry 52, 97-108, 618 doi:10.1016/j.apgeochem.2014.11.012. 619 Kamber, B.S., Greig, A., Schoenberg, R., Collerson, K.D., 2003. A refined solution to Earth’s hidden 620 niobium: implications for evolution of continental crust and mode of core formation. Precambrian 621 Research 126, 289-308, doi:https://doi.org/10.1016/S0301-9268(03)00100-1. 622 Kanematsu, M., Perdrial, N., Um, W., Chorover, J., O’Day, P.A., 2014. Influence of Phosphate and 623 Silica on U(VI) Precipitation from Acidic and Neutralized Wastewaters. Environmental Science & 624 Technology 48, 6097-6106, doi:10.1021/es4056559. 625 Kaplan, D.I., Xu, C., Huang, S., Lin, Y., Tolić, N., Roscioli-Johnson, K.M., Santschi, P.H., Jaffé, P.R., 2016. 626 Unique Organic Matter and Microbial Properties in the Rhizosphere of a Wetland Soil. 627 Environmental Science & Technology 50, 4169-4177, doi:10.1021/acs.est.5b05165. 628 Landais, P., 1996. Organic geochemistry of sedimentary uranium ore deposits. Ore Geology Reviews 629 11, 33-51, doi:10.1016/0169-1368(95)00014-3.

630 Langmuir, D., 1978. Uranium solution-mineral equilibria at low temperatures with applications to 631 sedimentary ore deposits. Geochimica et Cosmochimica Acta 42, 547-569, doi:10.1016/0016- 632 7037(78)90001-7. 633 Latta, D.E., Mishra, B., Cook, R.E., Kemner, K.M., Boyanov, M.I., 2014. Stable U(IV) Complexes Form 634 at High-Affinity Mineral Surface Sites. Environmental Science & Technology 48, 1683-1691, 635 doi:10.1021/es4047389. 636 Law, G.T.W., Geissler, A., Burke, I.T., Livens, F.R., Lloyd, J.R., McBeth, J.M., Morris, K., 2011. Uranium 637 Redox Cycling in Sediment and Biomineral Systems. Geomicrobiology Journal 28, 497-506, 638 doi:10.1080/01490451.2010.512033. 639 Lenhart, J.J., Figueroa, L.A., Honeyman, B.D., Kaneko, D., 1997. Modeling the adsorption of U(VI) 640 onto animal chitin using coupled mass transfer and surface complexation. Colloids and Surfaces A: 641 Physicochemical and Engineering Aspects 120, 243-254, doi:10.1016/S0927-7757(96)03865-4. 642 Li, K., Etschmann, B., Rae, N., Reith, F., Ryan, C.G., Kirkham, R., Howard, D., Rosa, D.R.N., Zammit, C., 643 Pring, A., Ngothai, Y., Hooker, A., Brugger, J., 2016. Ore Petrography Using Megapixel X-Ray Imaging: 644 Rapid Insights into Element Distribution and Mobilization in Complex Pt and U-Ge-Cu Ores. Economic 645 Geology 111, 487-501, doi:10.2113/econgeo.111.2.487. 646 Lidman, F., Mörth, C.M., Laudon, H., 2012. Landscape control of uranium and thorium in boreal 647 streams – spatiotemporal variability and the role of wetlands. Biogeosciences 9, 4773-4785, 648 doi:10.5194/bg-9-4773-2012. 649 Luo, W., Gu, B., 2008. Dissolution and Mobilization of Uranium in a Reduced Sediment by Natural 650 Humic Substances under Anaerobic Conditions. Environmental Science & Technology 43, 152-156, 651 doi:10.1021/es8013979. 652 Mehta, V.S., Maillot, F., Wang, Z., Catalano, J.G., Giammar, D.E., 2014. Effect of co-solutes on the 653 products and solubility of uranium(VI) precipitated with phosphate. Chemical Geology 364, 66-75, 654 doi:10.1016/j.chemgeo.2013.12.002. 655 Meunier, J.D., Trouiller, A., Brulhert, J., Pagel, M., 1989. Uranium and organic matter in a 656 paleodeltaic environment; the Coutras Deposit (Gironde, France). Economic Geology 84, 1541-1556. 657 Mikutta, C., Langner, P., Bargar, J.R., Kretzschmar, R., 2016. Tetra- and Hexavalent Uranium Forms 658 Bidentate-Mononuclear Complexes with Particulate Organic Matter in a Naturally Uranium-Enriched 659 Peatland. Environmental Science & Technology 50, 10465–10475, doi:10.1021/acs.est.6b03688. 660 Min, M.Z., Meng, Z.W., Sheng, G.Y., Min, Y.S., Liu, X., 2000. Organic geochemistry of paleokarst- 661 hosted uranium deposits, South China. Journal of Geochemical Exploration 68, 211-229, 662 doi:10.1016/s0375-6742(99)00085-0. 663 Newsome, L., Morris, K., Lloyd, J.R., 2014. The biogeochemistry and bioremediation of uranium and 664 other priority radionuclides. Chemical Geology 363, 164-184, doi:10.1016/j.chemgeo.2013.10.034. 665 Och, L.M., Müller, B., März, C., Wichser, A., Vologina, E.G., Sturm, M., 2016. Elevated uranium 666 concentrations in Lake Baikal sediments: Burial and early diagenesis. Chemical Geology 441, 92-105, 667 doi:10.1016/j.chemgeo.2016.08.001. 668 Owen, D.E., Otton, J.K., 1995. Mountain wetlands: Efficient uranium filters — potential impacts. 669 Ecological Engineering 5, 77-93, doi:10.1016/0925-8574(95)00013-9. 670 Paterson, D., de Jonge, M.D., Howard, D.L., Lewis, W., McKinlay, J., Starritt, A., Kusel, M., Ryan, C.G., 671 Kirkham, R., Moorhead, G., Siddons, D.P., 2011. The X-ray Fluorescence Microscopy Beamline at the 672 Australian Synchrotron. In: McNulty, I., Eyberger, C., Lai, B. (Eds.). 10th International Conference on 673 X-Ray Microscopy, pp. 219-222. 674 Ravel, B., Newville, M., 2005. ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption 675 spectroscopy using IFEFFIT. Journal of Synchrotron Radiation 12, 537-541, 676 doi:10.1107/s0909049505012719. 677 Read, D., Bennett, D.G., Hooker, P.J., Ivanovich, M., Longworth, G., Milodowski, A.E., Noy, D.J., 1993. 678 The migration of uranium into peat-rich soils at Broubster, Caithness, Scotland, U.K. Journal Of 679 Contaminant Hydrology 13, 291-308, doi:10.1016/0169-7722(93)90067-3.

680 Regenspurg, S., Margot-Roquier, C., Harfouche, M., Froidevaux, P., Steinmann, P., Junier, P., Bernier- 681 Latmani, R., 2010. Speciation of naturally-accumulated uranium in an organic-rich soil of an alpine 682 region (Switzerland). Geochimica et Cosmochimica Acta 74, 2082-2098, 683 doi:10.1016/j.gca.2010.01.007. 684 Rehr, J.J., Kas, J.J., Vila, F.D., Prange, M.P., Jorissen, K., 2010. Parameter-free calculations of X-ray 685 spectra with FEFF9. Physical Chemistry Chemical Physics 12, 5503-5513, doi:10.1039/b926434e. 686 Renock, D., Mueller, M., Yuan, K., Ewing, R.C., Becker, U., 2013. The energetics and kinetics of uranyl 687 reduction on pyrite, hematite, and magnetite surfaces: A powder microelectrode study. Geochimica 688 et Cosmochimica Acta 118, 56-71, doi:10.1016/j.gca.2013.04.019. 689 Rowell, D.L., 1994. Soil Science Methods and Applications. Longman UK, Harlow, England. 690 Ryan, C., 2000. Quantitative trace element imaging using PIXE and the nuclear microprobe. 691 International Journal of Imaging Systems and Technology 11, 219-230. 692 Ryan, C., Jamieson, D., 1993. Dynamic analysis: on-line quantitative PIXE microanalysis and its use in 693 overlap-resolved elemental mapping. Nuclear Instruments and Methods in Physics Research Section 694 B: Beam Interactions with Materials and Atoms 77, 203-214. 695 Ryan, C.G., Siddons, D.P., Kirkham, R., Li, Z.Y., Jonge, M.D.d., Paterson, D.J., Kuczewski, A., Howard, 696 D.L., Dunn, P.A., Falkenberg, G., Boesenberg, U., Geronimo, G.D., Fisher, L.A., Halfpenny, A., Lintern, 697 M.J., Lombi, E., Dyl, K.A., Jensen, M., Moorhead, G.F., Cleverley, J.S., Hough, R.M., Godel, B., Barnes, 698 S.J., James, S.A., Spiers, K.M., Alfeld, M., Wellenreuther, G., Vukmanovic, Z., Borg, S., 2014. Maia X- 699 ray fluorescence imaging: Capturing detail in complex natural samples. Journal of Physics: 700 Conference Series 499, 012002. 701 Shanbhag, P.M., Choppin, G.R., 1981. Binding of uranyl by humic acid. Journal of Inorganic and 702 Nuclear Chemistry 43, 3369-3372, doi:10.1016/0022-1902(81)80117-0. 703 Spirakis, C.S., 1996. The roles of organic matter in the formation of uranium deposits in sedimentary 704 rocks. Ore Geology Reviews 11, 53-69, doi:10.1016/0169-1368(95)00015-1. 705 Tinnacher, R.M., Nico, P.S., Davis, J.A., Honeyman, B.D., 2013. Effects of Fulvic Acid on Uranium(VI) 706 Sorption Kinetics. Environmental Science & Technology 47, 6214-6222, doi:10.1021/es304677c. 707 Tokunaga, T.K., Kim, Y., Wan, J., Yang, L., 2012. Aqueous Uranium(VI) Concentrations Controlled by 708 Calcium Uranyl Vanadate Precipitates. Environmental Science & Technology 46, 7471-7477, 709 doi:10.1021/es300925u. 710 Tokunaga, T.K., Wan, J.M., Pena, J., Brodie, E.L., Firestone, M.K., Hazen, T.C., Sutton, S.R., Lanzirotti, 711 A., Newville, M., 2005. Uranium reduction in sediments under diffusion-limited transport of organic 712 carbon. Environmental Science & Technology 39, 7077-7083, doi:10.1021/es050221a. 713 Tsarev, S., Collins, R.N., Fahy, A., Waite, T.D., 2016. Reduced Uranium Phases Produced from 714 Anaerobic Reaction with Nanoscale Zerovalent Iron. Environmental Science & Technology 50, 2595- 715 2601, doi:10.1021/acs.est.5b06160. 716 Vimy, 2014. Mulga Rock Uranium Project Resource Upgrade. Perth. 717 Vimy, 2015. Pre-Feasibility Study Reaffirms Mulga Rock Project as one of Australia’s Leading 718 Undeveloped Uranium Projects 719 Vimy, 2016a. Resource Update for Mulga Rock Project. 720 Vimy, 2016b. Significant Resource Upgrade for Mulga Rock Project. 721 Vimy, 2016c. Vimy Resources Limited; Annual Report 2016. Perth, p. 98. 722 Warwick, P., Evans, N., Hall, A., Walker, G., Steigleder, E., 2005. Stability constants of U(VI) and U(IV)- 723 humic acid complexes. Journal of Radioanalytical and Nuclear Chemistry 266, 179-190, 724 doi:10.1007/s10967-005-0890-7. 725 Wood, S.A., 1996. The role of humic substances in the transport and fixation of metals of economic 726 interest (Au, Pt, Pd, U, V). Ore Geology Reviews 11, 1-31, doi:10.1016/0169-1368(95)00013-5. 727 Wülser, P.-A., Brugger, J., Foden, J., Pfeifer, H.-R., 2011. The Sandstone-Hosted Beverley Uranium 728 Deposit, Lake Frome Basin, South Australia: Mineralogy, Geochemistry, and a Time-Constrained 729 Model for Its Genesis. Economic Geology 106, 835-867, doi:10.2113/econgeo.106.5.835.

730 Zammit, C.M., Brugger, J., Southam, G., Reith, F., 2014. In situ recovery of uranium — the microbial 731 influence. Hydrometallurgy 150, 236-244, doi:10.1016/j.hydromet.2014.06.003. 732 Zhao, D., Wang, X., Yang, S., Guo, Z., Sheng, G., 2012. Impact of water quality parameters on the 733 sorption of U(VI) onto hematite. Journal of Environmental Radioactivity 103, 20-29, 734 doi:10.1016/j.jenvrad.2011.08.010. 735 Zhou, P., Gu, B.H., 2005. Extraction of oxidized and reduced forms of uranium from contaminated 736 soils: Effects of carbonate concentration and pH. Environmental Science & Technology 39, 4435- 737 4440, doi:10.1021/es0483443. 738 739 740

1 Characterization of uranium redox state in organic-rich

2 Eocene sediments

3 Susan A Cumberland1,2,3, Barbara Etschmann2, Joël Brugger2, Grant Douglas4, Katy Evans5,

4 Louise Fisher6, Peter Kappen3, John W. Moreau1

5 1 School of Earth Sciences, University of Melbourne, Parkville, Victoria 3100, Australia

6 2 School of Earth, Atmosphere and Environment, Monash University, Clayton 3800, Victoria, 7 Australia

8 3 ANSTO Australian Synchrotron, 800 Blackburn Road, Clayton 3168, Victoria, Australia

9 4 CSIRO Land and Water, Floreat, Western Australia, Australia

10 5 Western Australian School of Mines, Curtin University, Bentley, Western Australia, Australia

11 6 CSIRO Mineral Resources, Bentley, Western Australia, Australia

Supplementary Information

14 SUPPLEMENTARY INFORMATION

15 SI Table 1 Location and drill depth of cores analysed

Core Depth Hole Id Deposit Samples from core Northing Easting RL Type (m)

NNA 5612 Princess MR5612a,b 6684044.69 578834.44 342.8 54 AC

NNA 5613 Princess MR5613a,b 6684001.28 578902.04 341.48 54 AC

NNA 5636 Princess MR5636 6684253.49 579486.8 345.52 72 AC

NNA 5706 Emperor MR5706a,b 6690894.00 552782.00 334.5 62 AC

NNA 5728 Shogun MR5728a,b 6686912.00 563587.00 318.4 51.0 AC

NNA 5766 Ambassador MR5766a,b,c,d 6682660.00 576969.00 342.1 64 AC

NND 5076 Emperor MR5076a,b 6682834.90 579131.20 339.5 48.5 DDH

NND 5077 Ambassador MR5077 6682189.32 576147.80 333 54.3 DDH

NND 5078 Ambassador MR5078 6682605.99 577075.96 342.4 57.0 DDH

CD1577 Ambassador 6682708.94 579942.81 331.4 43.5 DDH

16 AC = air core; DDH = diamond drill hole, RL = Reduced Level – elevation above sea level in metres of 17 the drill hole collar (Vimy, 2014, 2016b) 18 19

Supplementary Information

20 SI Table 2 Uranium mineral standards and location where known

Number Name Formula Location

M 33701 andersonite Na2CaUO2(CO3)3 · 6H2O

M 21676 autunite Ca(UO2)2 (PO4)2

M 30851 boltwoodite (Na, K)(UO2)2(HSiO4) · H2O

M 33175 brannerite U TiO4

M 32495 coffinite USiO4 USA, New Mexico Gransted, McKinley Country, Paddy M 28752 coffinite USiO4 Murphy Mines (35 8 50N, 107 51 3 W)

M 33214 curienite Pb2(UO2)2(V2O8)2 · 5H2O

M 22869 davidite (La, Ce)(Y, U, Fe)(Ti,Fe3+)20 (O,HO)38

M 45514 saleeite Mg(UO2)2(PO4)2 · 10H2O Democratic Republic of Congo (Zaire), Shaba (Katanga) M 44367 schoepite (UO2)O2(OH)12 · 12H2O Province, Kolwezi Munsonoi Mine (10 45 S 25 25 E)

M 17631 torbernite Cu(UO2)2(PO4)2 · 8 - 12 H2O Democratic Republic of Congo (Zaire), Shaba (Katanga) M 27299 soddyite (UO2)SiO4(H2O)2 Province, Swamba

M 26415 uraninite UO2

M 19266 uraninite UO2 Australia, Northern Territory, South Alligator River

M 34470 uranophane Ca(UO2)2(SiO3)(OH)2 · 5H2O

M 40774 uranopilite (UO2)6 (SO4)O2 (OH)6 (H2O)6 El Sherana mine, South Alligator River, NT, AU (13 31 S M 34904 uranopilite (UO2)6 (SO4)O2 (OH)6 (H2O)6 132 31 E)

M 25802 zippeite Mg (UO2)6 (SO4)3 (OH)10 (H2 O)n 21 Source: Museum Victoria, formulas are generic

Supplementary Information

24 SI Table 3 EXAFS fitting and paths for the uranyl nitrate standard

2 2 2 Ligand N R (Å) ss (Å ) ΔE0 Red X k-range R-range k-weighting SO 2-12.5 0.8 O1 2 (fix) 1.787(7) 0.002(1) 10(1) 978 1-4.5 Å-1 1,2,3 Å 5 O2 2 (fix) 2.34(3) 0.008(7)

N1 2 (fix) 2.47(4) 0.003 (fix)

O3 2 (fix) 2.53(3) 0.003 (fix)

O4_2(path 37) 4 (fix) 4.30(3) 0.003 (fix)

MS paths 21: O1-O1 3.57(1) 0.002(1)

22: O1-O1 3.57(1) 0.002(1)

24: N1-O1 3.65(4) 0.0025

27: O3-O1 3.71(3) 0.0025

error (quadrature) for MS paths

0.009899495

0.040607881

0.030805844

Supplementary Information

27 SI Table 4 Summary of physicochemical data in MR cores n = 18.

Carbonate - TOC LOI LOI Total N U (mg kg 550 900 % Sample ID Core Depth (m) Deposit 1 pH % % % % ) (**) CD1577 1577 43.5 Amb nd nd nd nd nd nd nd MR5076a 5076 44.3-44.4 Amb 1877 5.24 39 NA NA NA 0.46 MR5076b 5076 44.7-44.8 Amb 14930 3.09 30 NA NA NA 0.42 MR5077 5077 42.2-42.5 Amb 5760 4.12 44 NA NA NA 0.37 MR5078 5078 51.5-51.5 Amb 3720 3.02 34 NA NA NA 0.27 MR5612a 5612 41-42 Pri 2 4.7 60 57.7 78.3 17.3 0.35 MR5612b 5612 43-44 Pri 3.2 4.34 3.8 NA NA NA 0.04 MR5612c 5612 45-46 Pri 65 3.22 3.3 NA NA NA 0.05 MR5613a* 5613 40-41 Pri 5816 5.57 33.5 43.4 55.9 17.1 0.23 MR5613b* 5613 41-43 Pri 4588 5.61 39.5 NA NA NA 0.23 MR5636 5636 55-56 Pri 278 4.96 3.2 NA NA NA 0.05 MR5706a 5706 42-43 Emp 115 3.15 22 19.1 30.6 15.7 0.33 MR5706b 5706 55-56 Emp 40 7.06 0.72 1 1 0.1 0.02 MR5728a 5728 28.5-29 Sho 5590 3.33 26 37.9 41.9 5.1 0.29 MR5728b 5728 34-34.5 Sho 113 3.64 3.6 3.3 3.5 0.2 0.03 MR5766a 5766 53-53.5 Amb 5012 4.35 15.1 19.4 28.9 13 0.18 MR5766b* 5766 55.5-56 Amb 623 5.43 16 21.5 24.6 4.2 0.1 MR5766c* 5766 56.5-57 Amb 1206 4.60 21 22.5 33.2 14.6 0.14 MR5766d 5766 58-58.5 Amb 621 4.72 13 10.5 12.6 2.9 0.09 Mean 2798 4.5 22.5 23.6 31.1 9 0.2

Min 1.9 3 0.7 1 1 0.1 0

Max 14932 7.1 58.8 57.7 78.3 17.3 0.5

SD 3805 1.1 16.8 17.9 23.6 7.1 0.1

Median 914.2 4.5 21.2 20.4 29.8 9.1 0.2 28 *EXAFS Data available 2- 29 **NB estimated CO3 was calculated by LOI900 – LOI550 x MW C/ 100 (Dean, 1974; Heiri et al., 2001). Sample ID is the lab number 30 given by Melbourne University. NA = not analysed

Supplementary Information

31 SI Table 5 Elemental data for cores (1/2) 32 Sample number Core Depth Deposit U (XRF) U ashed U As Be Ca Co Cr Cu m mg kg-1 mg kg-1 mg kg-1 mg kg-1 mg kg-1 mg kg-1 mg kg-1 mg kg-1 mg kg-1

MR5076a 5076 44.3-44.4 Amb NA ND 1880 17 16 12900 282 261 624 MR5076b 5076 44.7-44.8 Amb NA ND 14900 296 128 5290 4300 163 7290 MR5077 5077 42.2-42.5 Amb NA ND 5760 49 3 2950 142 880 458 MR5078 5078 51.5-51.5 Amb NA ND 3720 157 19 34 541 337 1500 MR5766a 5766 53-53.5 Amb 5280 5198.0 5010 50 20 2450 2960 535 696 MR5766b 5766 55.5-56 Amb 642 586.4 623 4 3 912 70 886

Supplementary Information

35 SI Table 5 Elemental data for cores (2/2) 36 sample number core depth deposit Ga Li Ni Pb Rb Sc Ti Th V W Zn total REE+Y m mg kg-1 mg kg-1 mg kg-1 mg kg-1 mg kg-1 mg kg-1 mg kg-1 mg kg-1 mg kg-1 mg kg-1 mg kg-1 mg kg-1

MR5076 5076 44.3-44.4 Amb 3 81 1250 158 1.9 192 5270 44 145 1.1 1750 6270 MR5076 5076 44.7-44.8 Amb

Supplementary Information

38 SI Table 6 Summary of metals within deposits shown as average data in parts per million (mg kg-1)

Mean Min Max Ambassador Emperor Princess Shogun Element n=8 n=2 n=6 n=2 As 45.2 0.3 295.7 72.3 3 36.7 4.6 Be 12.2 0.1 128.3 24.6 0.8 3.3 0.3 Ca 2553 33.6 12950 3260 466.9 2966 569.9 Co 614.8 0.5 4300 1161 78.2 270.1 0.6 Cr 621 19.4 2552 668.8 355.7 662.7 570.1 Cu 1696 0.9 7286 2325 30.1 2241 111.8 Ga 9.6 0 48.8 8.3 10.1 6 24.8 Li 13.4 0.9 80.5 19.2 13.1 8.8 4.1 Ni 1594 0.7 10800 3128 54.5 591.1 2.5 Pb 1033 3.7 5292 1690 35.5 754.7 235.9 Rb 2.7 0.3 11.8 2.4 6.1 2 2.2 Sc 314.4 6.8 1822 500 15.7 207.3 191.9 Th 42.6 1.1 244.1 24 12.4 44.4 141.6 Ti 8934 474.3 45450 13290 2720 6516 4997 U 2798 1.9 14930 4219 77.9 1792 2853 V 139 4.1 492.7 205.7 51.1 98.1 82.7 W 3.2 0.6 9.2 3.6 2.4 3 3.2 Zn 2673 10.1 12770 3225 19.6 3382 12.8 REE+Y 2168 14.6 15390 4274 156.8 665.3 264.5 39

Supplementary Information

41 SI Table 7 Correlations between U and metals; TOC and metals

Correlation Test R R2 p-value significance U vs TOC 0.377 0.142 0.1225 * U vs As 0.861 0.741 0.0000 *** U vs REE+Y 0.805 0.649 0.0001 *** U vs Ba 0.14 0.02 0.5786 ns U vs Ca 0.345 0.119 0.1606 * U vs Cd 0.816 0.666 0.0000 *** U vs Mo 0.195 0.038 0.4377 * U vs Ni 0.602 0.363 0.0082 ** U vs Pb 0.875 0.765 0.0000 *** U vs Sn 0.234 0.055 0.3491 * U vs Sr 0.278 0.077 0.2641 * U vs Th 0.282 0.08 0.2567 * U vs V 0.577 0.333 0.0121 ** U vs Zr 0.113 0.013 0.6548 ns

TOC vs As 0.344 0.118 0.1626 * TOC vs REE+Y 0.258 0.067 0.3006 * TOC vs Ba 0.302 0.091 0.2233 * TOC vs Ca 0.591 0.35 0.0097 ** TOC vs Cd 0.145 0.021 0.5663 ns TOC vs Mo -0.069 0.005 0.7841 ns TOC vs Ni -0.01 0 0.9695 ns TOC vs Pb 0.457 0.208 0.0568 * TOC vs Sn 0.349 0.122 0.1552 * TOC vs Sr 0.54 0.291 0.0208 ** TOC vs Th 0.184 0.034 0.4643 * TOC vs V 0.481 0.231 0.0433 ** TOC vs Zr 0.021 0 0.9349 ns 42 *** p<0.005, ** p<0.05, * p <0.5, ns p >0.51

Supplementary Information

45 SI Table 8 Results from the linear combination fitting. Two end-members were represented by synthetic uraninite (U(IV)) and mineral uranopilite (U(VI))

uraninite Uranopilite Data Core Rfactor Chinu chisqr nvarys scaleby weight error weight error MR5766c NNA 5766 0.0004 9.89E-05 0.017 1 1 0.14 0.0043 0.86 0.0043 MR5766d NNA 5766 0.011 0.0027 0.452 1 1 0.35 0.0226 0.65 0.0226 MR5713b NNA 5613 0.008 0.0019 0.316 1 1 0.36 0.0188 0.64 0.0188 MR5713a NNA 5613 0.0005 0.00013 0.022 1 1 0.12 0.0049 0.88 0.0049 MR5076a NND 5076 0.006 0.0012 0.204 1 1 0.33 0.0151 0.67 0.0151 MR5077 NND 5077 0.007 0.0016 0.266 1 1 0.24 0.0173 0.76 0.0173 MR5076b NND 5076 0.002 0.0004 0.072 1 1 0.19 0.0090 0.81 0.0090 MR5766b NNA 5766 0.003 0.0008 0.129 1 1 0.11 0.0120 0.89 0.0120 reduced_CD1577 CD 1577 0.015 0.0025 0.419 1 1 0.65 0.0217 0.35 0.0217 oxidised_CD1577 CD 1577 0.003 0.0005 0.091 1 1 0.16 0.0101 0.84 0.0101 U(IV) Synthetic uraninite (Tsarev et al., 2016) 1 U(VI) Natural uranopilite M 40774 1 46

Supplementary Information

48 SI Table 9 Alternative EXAFS model using more MS paths

k-range 2 2 2 Sample Ligand N R (Å) ss (Å ) ΔE0 Red X r-factor (Å) R-range (Å) k-weighting S0 MR5766c Oax2 1 1.75(4) 0.003 (fix) 7(1) 111 0.025 2 - 12 1.3 - 4 1,2,3 0.85 Oax1 1 1.77(3) 0.003 (fix) O15 2 2.20(1) 0.003 (fix) O14 2 2.34(2) 0.003 (fix) OH20 2 2.39(2) 0.003 (fix) C 2 2.92(3) 0.003 (fix) U-Oax2-U-Oax2 (MS) 3.51(6) 0.006 (fix) U-Oax2-Oax1 (MS) 3.58(5) 0.006 (fix) U-Oax2-U-Oax1 (MS) 3.58(5) 0.006 (fix)

MR5713a Oax2 1 1.77(10) 0.003 (fix) 9(3) 111 0.025 3 - 11.5 1.3 - 4 1,2,3 0.85 Oax1 1 1.77(8) 0.003 (fix) O15 2 2.23(5) 0.003 (fix) O14 2 2.38(4) 0.003 (fix) OH20 2 2.43(6) 0.003 (fix) C 2 2.89(7) 0.003 (fix) U-Oax2-U-Oax2 (MS) 3.54(11) 0.006 (fix) U-Oax2-Oax1 (MS) 3.58(13) 0.006 (fix) U-Oax2-U-Oax1 (MS) 3.58(13) 0.006 (fix)

oxidised_CD1577 * Oax2 1 1.75(4) 0.003 (fix) 7(1) 111 0.025 2.5 - 12 1 - 4 1,2,3 0.85 Oax1 1 1.77(3) 0.003 (fix) O15 2.8(4) 2.20(1) 0.003 (fix) O14 2 2.34(2) 0.003 (fix) OH20 2 2.39(2) 0.003 (fix) C 2 2.88(4) 0.003 (fix) U-Oax2-U-Oax2 (MS) 3.50(6) 0.006 (fix) U-Oax2-Oax1 (MS) 3.58(5) 0.006 (fix) U-Oax2-U-Oax1 (MS) 3.58(5) 0.006 (fix) 49

50 *these distances were constrained to be the same as those of MR5766c these 3 data sets were fitted together.

Supplementary Information

51 SI Figure 1 XRD spectra

52 A) Uranyl nitrate Patterns were collected with a Bruker D8 Advance Eco diffractometer equipped with a

53 LYNXEYE XE linear position sensitive detector and using a Cu X-ray tube operated at 40 kV and 25 mA.

54 Data were collected from 3–80° 2θ with a step size of 0.02°2θ and a counting rate of 1 s/step. Minerals

55 were identified with reference to standard patterns from the ICDD Powder Diffraction File 2 (PDF-2)

56 database using the DIFFRAC.EVA v.4 software package (available from Bruker AXS).

Supplementary Information

65 B) Background of CD1577. Data were measured using a Bruker GADDS microdiffractometer, using Cu Kα

66 radiation from an X-ray tube operating at 40 mA, 40 kV. The incident beam was passed through crossed

67 reflecting (Gobel) mirrors, resulting in a monochromatic, parallel, high-brilliance source which was then

68 collimated through a 200 µm pinhole system. The spectrum was consistent with quartz alpha mixed with

69 an unidentified minor phase.

Supplementary Information

76 C) Coffinite- Micro-XRD pattern of mineralised area of uranium, spot analysis was taken at the edge of a

77 pyrite grain. Red lines show markers corresponding to U(IV) coffinite pattern (USiO4) (RRUFF). Beam =

78 300 µm Cu Kα collimator.

Supplementary Information

82 SI Figure 2 U L3 edge XAS x-ray absorption of the near edge structure for spectra of bulk cores and of some 83 mineral standards

86 SI Figure 3 U L3 edge XAS - Extended x-ray absorbance of the fine structure shown in K2 space. Top four 87 spectra are fitted EXAFS of the bulk cores where solid line is the sample. The U(IV) and U(VI) lignite samples 88 were from a thin section of MR lignite (CD 1577). The bottom seven spectra are mineral standards donated by 89 Museum Victoria.

Supplementary Information

92 SI Figure 4 EDS-SEM map and spectra from CD1577

94 95